can be obtained only when a collection vessel is used before measurement. Once formed, the hydrides
are entrained in argon and swept into the light path where they are decomposed to produce an elemental
vapour. The decomposition is achieved either by direct injection into a 'cool' hydrogen flame or in a
long path (15 cm) silica cell with external heating. Both air/acetylene flames and electrothermal heating
have been used.
Recent developments in the determination of elements in this group have been very much linked to the
use of atomic fluorescence detection systems rather than AAS (see section 8.7). ICP-AES and ICP-MS
can also be used but they are generally inferior in sensitivity. Best sensitivity is obtained from AFS
detection. It should also be noted that the analysis may also be required to detect and measure organic
compounds of these elements because of the toxicity in the organic form. Separation by one of the
methods reviewed in Chapter 4 may thus be used in sample processing prior to analysis.
Quantitative Measurements and Interferences
Quantitative measurements may be made using a previously prepared calibration curve or by the
method of standard addition. In either case, operating conditions must first be optimized with regard to
the expected concentration range of samples and linearity of response. This involves appropriate choice
of resonance line (usually made by reference to tables), adjustment of lamp current, flame temperature
and sample aspiration rate, burner alignment and monochromator slit width. Standard solutions are best
prepared by appropriate dilution of 1000 ppm stock solutions and should be matched as closely as
possible in gross composition to those of the samples. The relative precision of atomic absorption
measurements is good, and in most cases 0.5–2% is attainable without difficulty where flame
atomization is used. Precisions for flameless methods are, however, often much poorer as a result of
some of the interferences discussed below. Calibration curves invariably show curvature towards the
concentration axis when absorbances exceed 1. This non-linearity is caused by unabsorbed radiation
reaching the detector or when the half-width of the emission line from the lamp approaches or exceeds
that of the absorption line (p 322 et seq.). Unabsorbed radiation may reach the detector from numerous
sources, including emission lines of the cathode material close to the chosen resonance line or from the
filling gas, scattered radiation in the monochromator, and radiation bypassing the flame or sample
vapour.
Interferences in atomic absorption measurements can arise from spectral, chemical and physical
sources. Spectral interference resulting from the overlap of absorption lines is rare because of the
simplicity of the absorption spectrum and the sharpness of the lines. However, broad band absorption
by molecular species can lead to significant background interference. Correction for this may be made
by matrix matching of samples and standards, or by use of a standard addition method (p. 30 et seq.).